ContextQMD
Back to Sentence Transformers

Evaluation

docs/package_reference/multi_vector_encoder/evaluation.md

6.0.01020 B
Original Source

Evaluation

sentence_transformers.multi_vector_encoder.evaluation defines evaluators tailored to multi-vector (late-interaction) models. All evaluators score with MaxSim and accept the multi-vector model's ragged per-token embeddings.

MultiVectorInformationRetrievalEvaluator

{eval-rst}
.. autoclass:: sentence_transformers.multi_vector_encoder.evaluation.MultiVectorInformationRetrievalEvaluator

MultiVectorNanoBEIREvaluator

{eval-rst}
.. autoclass:: sentence_transformers.multi_vector_encoder.evaluation.MultiVectorNanoBEIREvaluator

MultiVectorTripletEvaluator

{eval-rst}
.. autoclass:: sentence_transformers.multi_vector_encoder.evaluation.MultiVectorTripletEvaluator

MultiVectorRerankingEvaluator

{eval-rst}
.. autoclass:: sentence_transformers.multi_vector_encoder.evaluation.MultiVectorRerankingEvaluator

MultiVectorDistillationEvaluator

{eval-rst}
.. autoclass:: sentence_transformers.multi_vector_encoder.evaluation.MultiVectorDistillationEvaluator